Developer based in Sweden, building testable, observable AI agents and agentic tooling in Python.
- agentkit: Python toolkit for building AI agents you can test, observe and bound: evals, tracing, budgets, tool approvals, MCP.
- ai-task-orchestrator: Run AI coding backlogs unattended with provider failover (Claude Code, Codex, Copilot), verification gates, and zero telemetry.
- agentic-ai-tutorials: Hands-on tutorials for agentic AI with Google ADK and LangChain/LangGraph, runnable on Gemini or local models.
- InvestingAwesomely-IAOS: Vendor-neutral, explainable investment operating system.
- crypto-paper-trading: Paper-trading simulator with live crypto market data (FastAPI + React).
AI agents · LLM evaluation · observability · developer tooling · Python

